Help With Setup

[nxs450]

Hello I am new here recommended by another member. I am having a problem with my system when I take measurements with REW. My center channel is either coming back 5db to 10db hotter then left, right , or left and right is 5db to 10db lower then center. Any ideas on what could cause this?

 

[John Mulcahy]

How is REW connected to the system?

 

[nxs450]

HDMI

 

[Wayne A. Pflughaupt]

What is the brand / model of the speakers?

Regards,
Wayne

 

[nxs450]

DIYSG HTM12's

 

[John Mulcahy]

Do the channels have the same level corrections in the processor? Any processing engaged? Some processors have a setting to boost speech clarity by raising the level of the centre channel.

 

[nxs450]

Do the channels have the same level corrections in the processor? Any processing engaged? Some processors have a setting to boost speech clarity by raising the level of the centre channel.

Yeah I am thinking that is what may be going on. It is a new Denon X4300 and it happens with or without Audyssey applied. Someone at another site thought that this was a issue but maybe designed that way.
What do you think my measurements disregarding the center being a little hotter?

 

[Matthew J Poes]

I'm sorry if I missed this, but the levels are all set equal and crossover settings are the same? I've never seen anything like that before myself.

Dialogue enhancement can be it, but I thought that was more than just an overall boost, that it was frequency specific.

There really isn't a lot of information to go by, but based solely on the frequency response at that one location, it looks very good. The exact bass tuning is really more up to you. The bass looks hot compared to the left and right and a bit too flat compared to the center, but that is a matter of taste. I know Matt Grant well and love the HTM-12's for the exact reason your measurements look so good. They are a well designed speaker and your measurements show typical in room response behavior for speakers of this pedigree. Those are great speakers, enjoy! I consider them to be a spiritual cousin to my own speakers which is one reason why I like them so much.

While you may not need it for absolute output, the HTM's can handle a lot more power than your receiver can deliver. They can benefit from an external amplifier which has the side benefit of reducing the strain on the receiver and freeing up more power for the other speakers.

 

[nxs450]

I'm sorry if I missed this, but the levels are all set equal and crossover settings are the same?

Yes Audyssey set them pretty much the same. Audyssey actually set the center 2db lower then left and right. It would be nice if someone else with Denon X4300 could check for the same.

I have been very happy with the HTM's. If you don't mind me asking what speakers are you running?

I am looking at amps on ebay all the time. Do you think these speakers will sound better from higher powered amp at moderate or close to reference levels?

 

[Matthew J Poes]

Yes Audyssey set them pretty much the same. Audyssey actually set the center 2db lower then left and right. It would be nice if someone else with Denon X4300 could check for the same.

I have been very happy with the HTM's. If you don't mind me asking what speakers are you running?

I am looking at amps on ebay all the time. Do you think these speakers will sound better from higher powered amp at moderate or close to reference levels?

I use Gedlee Abbey's which are a 12" waveguide design from Earl Geddes. They are conceptually similar using B&C DE250's (which the HTM's use a Chinese made improved clone) and 12TBX100. The HTM's use an eminence driver which isn't as expensive or have as fancy design features, but in practice its an amazing driver. It has very smooth response and very low distortion. The biggest advantage of the B&C 12 that mine use is a lot more power handling and xmax, but neither speaker has much bass so this isn't of practical importance. The power handling might be, but your speaker could handle over 500 watts and while mine can technically handle over 1000, at that point the crossovers become a weak link. Again, not of great practical importance. The HTM uses the SEOS waveguide which intentionally took advantage of Earl Geddes work to optimize its design, and arguably improve upon it. Hence why I consider the HTM to be a spiritual cousin. We aren't just talking about a 12" waveguide two-way with similar drivers, the waveguides themselves are similar.

At moderate volume when a solid-state amplifier is not clipping, you won't hear much difference if any at all. It's really at high volumes when the speakers are drawing enough power to clip. This is where the more powerful amplifiers will make a difference. While these speakers are very efficient, you can still find in many rooms where getting up at or slightly above reference will clip the amplifiers. My center channel was running on my 100+ watt receiver, my speakers are 96db's at 1 watt 1 meter, and I was clipping my receiver below reference levels while running them full range. Adding a high pass filter prevented this until much more uncomfortable levels, but it was still clipping during some dynamic peaks. This isn't surprising actually. Plug my 96db efficiency and 12 foot listening distance into a peak SPL calculator and it comes up with 104.7db's. That means I barely met the criteria for reference levels under that scenario. The calculator is imperfect and most receivers won't deliver their rated power during dynamic peaks because other speakers (In my case the 4 surround speakers) are also drawing from the receiver. This all explains why I was able to clip the amp.

Be creative when looking up amps, you might find better values. It needs to be 2-4 times more powerful than your receiver. A 200 watt per channel or 400 watt per channel amplifier would be good. Each doubling will give you just 3db's more headroom/output. Beyond that, remember that as long as we are dealing with a minimum quality of amplifier here that isn't too old, they won't sound much different. A mid 90's or early 2000's Parasound, Adcom, Anthem, etc. will all work great. A recent Emotiva, an ATI or any amp made by ATI, etc.

 

[nxs450]

Okay I figured out why the center channel was measuring so hot. I had checked dialog level adjust and it was set at 0db, no gain, nor defeat, but just the fact it was turned on it was boosting the center. I had assumed that since it was set at 0db it wasn't causing the problem.

New measurements with dialog level off and some other changes.

I have a question about REW. The reason I purchased the Microphone and started Measuring was to make sure there were no major problems with my room. I also see others on another site that are posting all kinds of different measurements. Now it seems OCD is kicking in and chasing perfection. lol However my understanding and knowledge of all this is pretty limited. My question is based on this measurement is it considered good enough, and if so are there other things I should check?

Alan

 

[Wayne A. Pflughaupt]

Excellent response! Impressive low end extension too – what sub are you using?

Regards,
Wayne

 

[Talley]

Wow... a tad hot on the bottom end isn't it?

This is mine at 1/6... Just trying to understand is all.

 

[Matthew J Poes]

Wow... a tad hot on the bottom end isn't it?

This is mine at 1/6... Just trying to understand is all.View attachment 5879

Assuming the db’s are correct you should measure at a lower volume. No need to measure that loud. You put your hearing and system at risk and sometimes see flaws show up that aren’t really there, just resonances in the room.

 

[nxs450]

Excellent response! Impressive low end extension too – what sub are you using?

Regards,
Wayne

Thanks! 2 DIY ported 460HO's driven by Inuke 6000 and 2 sealed PA460's driven by Inuke 3000

 

[Matthew J Poes]

Okay I figured out why the center channel was measuring so hot. I had checked dialog level adjust and it was set at 0db, no gain, nor defeat, but just the fact it was turned on it was boosting the center. I had assumed that since it was set at 0db it wasn't causing the problem.

New measurements with dialog level off and some other changes.

I have a question about REW. The reason I purchased the Microphone and started Measuring was to make sure there were no major problems with my room. I also see others on another site that are posting all kinds of different measurements. Now it seems OCD is kicking in and chasing perfection. lol However my understanding and knowledge of all this is pretty limited. My question is based on this measurement is it considered good enough, and if so are there other things I should check?

Alan

That’s a great response. You definitely have the bass hot but if you like it that’s ok. I would also measure at a much lower volume in the future. Something like 75dbs on average is more than adequate. Your bass is so hot that you are likely pushing the low end limits to some extent. Even with a ton of uber subs you get to a point where they may be good for another 20-30dbs but distortion is already rising. You don’t want to measure that this way, your primary interest is just response.

The number one concern in most systems is low frequency linear distortion. That is, an uneven anomalous response in the bass. You have as flat of a bass response as one can expect.

As for what else to look at. The most people important thing you want is a flat response over a range of mic positions. You may think that only one position matters because you only sit in one spot. That actually misses the point. Your head won’t be exactly where the mic is and your ears won’t hear as the mic does. As such you want to take a set of measurements spaces 1-2 feet apart at least. If you want smooth bass over all your seating positions you want to measure in those too. When you do this you should label the position carefully. If any have a weird response I would suggest labeling their distance to boundaries as sometimes an extreme position such as a seat near a wall will show a weird response. Nothing you can do about that. Sucks for whoever sits there.

It’s not hard to get a flat response at a single mic position. The problem is that doing his can actually increase the response variance over a range of mic positions and those mic positions may still be all representative of your head position. Thus that is the real trick, eqing a response so it’s flat generally where your ears will be.

As for other things to look at, while not as important I do look at the following:

1. RT60 and EDT (should fall between .2 and .5, lower is common in smaller rooms), you want it flat down to 200hz. A rise below 200hz is fine. It isn’t a valid measure below Schroeder.
2. Distortion: I look to see that there is no unusual rise in distortion. I didn’t used to feel this had any value but have recently found a number of well regarded speakers with a rise in distortion at average listening levels suggesting the tweeter crossover is too low or the tweeter itself isn’t great.
3. Impulse response: I look at the various reflections and their strength. I sometimes window the measurement to look at the response of the reflection. You want them to be lower in level than the original impulse but of the same basic response shape. If they have a bad looking response you want them to be down at least 10-15dbs to ensure they aren’t overly audible.

Things I really don’t find to be important but are common on the forums:

1. CSD plot: ok so because bass is minimum phase a flat response has a flat decay. A non-flat response has a non-flat decay. While the overall rate of decay could be increased with traps, this is hard to measure with a CSD. Hence I’ve largely stopped interpreting CSD. I find one and time again that the minimum phase concept holds true and if it doesn’t I can always point to noise floor issues and not non-even decay. There are some “experts” who make a very big deal about CSD. These experts are inaccurate in their views as measurements and theory both prove.
2. Spatially averaged RTA measurements. Now there are nothing wrong with or invalid with these. However I don’t find them to give me anything that a set of normal measurements won’t give me and I have concern handling the mic. I have found problems in these measurements as a result of how I handle the mic. I don’t see this as having value. It’s only real advantage is that a spatially averaged response is more quickly obtained. However it won’t be repeatable and like I said, may be inaccurate.
3. Any detailed review of distortion. It’s a good guy check to be sure nothing is wrong, but high distortion can be as much a function of something in the room as it is of a speaker flaw. I’ve also had blips in the distortion plots as a result of non-coninupus ambient noise such as talking or walking sounds. I recently took measurements at an audio showroom and got really high bass distortion. They had another demo room playing with the door closed. I turned that demo room off and retook and the bass distortion went away. I took measurements at my fathers house his past week and my dad kept talking during the measurement. I got blips in the distortion curve (and this is where I saw a strange rise in the tweeter distortion). I retook then after everyone left for a while and the blips went away, but the tweeter distortion remained. My point here is that it’s worth a quick check but don’t put too much weight in them and I would certainly not compare them to other people’s measurements.
4. Clarity: I actually believe clarity has merit in a theater but here is the state of the Union right now: zero research has ever examined clarity measures in a small room. There is no knowledge of what to look for or what is good. It’s just a number. On top of that, hey are easy to screw up and will almost always be very high on small absorbent rooms. I would love to research this and establish values but as of right now there is nothing you can really glean from clarity measures in a home theater.

 

[Matthew J Poes]

Wow... a tad hot on the bottom end isn't it?

This is mine at 1/6... Just trying to understand is all.View attachment 5879

Your measurements are excellent as well.

The shape of your response makes me think this is Dirac right? It’s hard to get that perfect wedge shape without something like Dirac. It’s possible with flexible eq but I rarely see anyone doing that. Most people only have flexible eq on the subs.

Research into good room curves suggest that the most preferred response shape has a rise from 20khz to 20hz at 1db per octave. That will mean that 20hz should be 10-12dbs hotter than 20khz. However that was an average based on people’s preferenecs and closely matched the natural response shape that an anechoicly flat speaker would show once placed in a typical room. Loudness curves help us understand that we are very insensitive to bass frequencies. In fact based solely on that we would likely prefer a dynamically eqed system that is a good 20dbs too high at average listening levels. As such it seems to be generally accepted that the absolute levels of bass can vary quite a bit and still be considered neutral. We know that flat is not neutral at all, but just how much of a slope is acceptable is unclear.

 

[Talley]

Assuming the db’s are correct you should measure at a lower volume. No need to measure that loud. You put your hearing and system at risk and sometimes see flaws show up that aren’t really there, just resonances in the room.

Yes, correct. However measuring while the family is home and your noise floor is around 50db I wanted to get over that. I also know that my room resonates from 24hz and lower but by 29hz and up there is no resonance.

I also was looking at distortions at those levels as well.

I normally will test at 85db.

 

[Matthew J Poes]

Yes, correct. However measuring while the family is home and your noise floor is around 50db I wanted to get over that. I also know that my room resonates from 24hz and lower but by 29hz and up there is no resonance.

I also was looking at distortions at those levels as well.

I normally will test at 85db.

You can still probably get away with 85dbs, but if you want to measure that high it’s fine.

What did the distortion look like?

 

[Talley]

You can still probably get away with 85dbs, but if you want to measure that high it’s fine.

What did the distortion look like?

I have no clue... THD seemed low? but not sure what I'm looking at with all the fundamentals. I've attached the file for review. It seems distortion is high around the crossover point.

I agree... I was pushing the system to it's limits at volume level of -5db

I would normally test at -15db.

 

[Matthew J Poes]

I’ll take a look and let you know. I’ll respond via pm so as not to hijack the thread. It will be a bit as I’m tied up at the moment.

 

[nxs450]

Wow... a tad hot on the bottom end isn't it?

This is mine at 1/6... Just trying to understand is all.View attachment 5879

Yeah it is hot on the bottom, I have Cinema EQ applied and it boosts it like that. With it off it is pretty much flat all the way across. Same with me I just started this a month ago from scratch, been doing a lot of reading. lol

 

[nxs450]

I have a ETC measurement where I am checking reflections. I don't know why it is reading negative numbers at a suspect point of -15 and 7ms that reads -9.62. In the REW guide it has an example reading at point of -15 and 5ms that reads 6.1ft.
EDIT: I figured it out I was dragging between two points and left to right which reads negative. Can ceilings cause the reflection on this measurement.

If I am reading the guide right I take the measurement I get off the graph at suspect point, add it to distance of speaker to microphone (that Audyssey set) and cut a string the length of the two measurements, secure one end at mic and one at speaker, then grasp string in center and move to a point where it touches a hard/untreated surface?

 

[Matthew J Poes]

I have a ETC measurement where I am checking reflections. I don't know why it is reading negative numbers at a suspect point of -15 and 7ms that reads -9.62. In the REW guide it has an example reading at point of -15 and 5ms that reads 6.1ft.
EDIT: I figured it out I was dragging between two points and left to right which reads negative. Can ceilings cause the reflection on this measurement.

If I am reading the guide right I take the measurement I get off the graph at suspect point, add it to distance of speaker to microphone (that Audyssey set) and cut a string the length of the two measurements, secure one end at mic and one at speaker, then grasp string in center and move to a point where it touches a hard/untreated surface?

Yes that's basically it. The distance is the difference is path length between mic and speaker and the reflection.

I've never used a string myself. I take the difference and then the path length of various reflections and see which matche the new Total distance.

It's all basically trigonometry. Don't let that scare you though, you don't need to figure out the math. The point is, If you know your speakers location and the distance between the speakers and walls, you can use trig to calculate the path length of all the reflections. You can then match these to the ETC plot to figure out which is which.

This all works fine in small rooms but not so well in large rooms. Large rooms often have tons of reflections that are similar in path length so it's hard to tell what is what. In a small room it's pretty easy to figure out where the reflections are coming from.

Yes ceiling and floor reflections typically show up in these plots. Even radiation out the back of the enclosure shows up in these. Many people don't realize that speaker enclosures do not contain the sound of the speaker all that well and that fairly loud sound is radiated out of an enclosure. As such you can see reflections from this sound coming out the back of the box. That can make identifying the source a little tricky. You have to consider every possible reflection.

You can also move the IR center to align with the center of each reflection. This let's you look at the response of the reflection (very roughly).

 

[nxs450]

Thanks Matt

So as per REW 101 the peaks above -20 should be addressed. All 3 of my front speakers exhibit almost the same peaks. I'm not seeing very many above -20, in the measurement that I posted, am I correct about that? I actually did the string trick last night and it worked. The peak around 16ms is on the back wall beside my equipment rack, I treated my side walls with 2" OC 703 but did not do the back. When I put a 2'x4' piece of acoustic material in the spot, the peak went away. Same with the peaks around 6ms to 8ms, which ended up on the ceiling above front row seating.

What about the peaks right at 0ms to 2ms, is that something that needs addressed?

Alan